Excavation and trenching are among the most hazardous operations managed by utilities.
AI is providing us with a new approach to flag security risks.
How can you help?
You can contribute to improving excavation safety by recording a short structured video using vyn. vyn will automatically label it, index and extract the relevant information and export to an open dataset.
60 seconds of your time will make a big difference to make excavations safer!
Every new single data element we add here helps build a better prediction.
What scenes are you looking for?
Look around you and simply capture, if you see any of the following situations:
- Access to trench areas where work is taking place
- Security signage - e.g. “Workers in the road” signs
- People wearing Personal Protective Equipment (PPE)
- Barriers, fences around the trench or dug-up area
- People working at a height, on ladders or a work platform
- People working in trenches on ladders
- Excavators or diggers
- Other security observations that could help someone remote to assess safety risk
Examples of images
We have selected a few images (marked as being in the public domain under Creative Commons license) to show you what safety scenes to look for and capture.
Workers wearing Safety Helmets
|"Obreros" by Daquella manera is marked with CC0 1.0||"USACE works to reduce flood risk in Napa, Calif." by USACE HQ is marked with CC PDM 1.0|
Workers working at Height
|"USACE contractors work at Prompton Dam and Reservoir" by USACE HQ is marked with CC PDM 1||"Workers install rebar at Prompton Dam and Reservoir" by USACE HQ is marked with CC PDM 1.0|
Cordon Barriers & Cones
|"20161016-FNS-LSC-0230"by USDAgov is marked with CC PDM 1.0||"Cuidado" by Daquella manera is marked with CC0 1.0|
At vyn, we believe that Artificial Intelligence (AI) will drive social and economic impact.
We also know that a vibrant eco-system in AI Research & Development is crucial to drive innovation and contribute to AI technology on the global AI map.
We want to drive collaboration between government, academic institutions, industries, startups and other players for improving the impact of research and converting it into outcomes.
We want to encourage young researchers to take on challenging problems and find new solutions.
And of-course, we all know that the biggest handicap in training deep learning and AI algorithms is the need for large volumes of clean data.
We recognise that there are many data repositories already in use by researchers and developers, but our own experience has shown us that most of these these focus on academic problems. Collecting data to address business problems in utilities, telecom, manufacturing continues to be a challenge.
We are making available to the research community our smart video tool (vyn) to collect structured and labelled data, and will make public any datasets that we generate through this initiative. We hope that our little initiative will augment existing efforts and build interest in researching business problems.